Why Q-chem does not consider FOLLOW_OVERLAP setting

Hello,

I am attempting to optimize an excited state by applying the following parameters:

STATE_FOLLOW        TRUE
FOLLOW_ENERGY       1000000
FOLLOW_OVERLAP      90

I chose this set to ensure that the state will not change during optimization. However, in the output, Q-Chem does not seem to consider FOLLOW_OVERLAP=90 and unexpectedly follows the energy as usual. In the output, it mentions that dynamic thresholds are being used to change the set FOLLOW_ENERGY and FOLLOW_OVERLAP. My question is, why does not Q-Chem consider my setting for FOLLOW_OVERLAP=90, and how can I fix this problem?

A part of output:

 Step:              1 State:            28
 Previous State: 28  -1578.671124720782
 Estimated State:    -1578.639342681677
 Energy threshold: 1.00000 Overlap threshold:  0.900
 ---------------------------------------------------
 State |    Energy (Eh)   | Delta E (Eh) |  Overlap 
 ---------------------------------------------------
  12   |   -1578.69154355 |  -0.052201   |  0.420332
  20   |   -1578.67887588 |  -0.039533   |  0.788062
  21   |   -1578.67768797 |  -0.038345   |  0.825723
  22   |   -1578.67486410 |  -0.035521   |  0.677087
  23   |   -1578.67468462 |  -0.035342   |  0.628466
  24   |   -1578.67420352 |  -0.034861   |  0.759117
  25   |   -1578.67287486 |  -0.033532   |  0.657837
  27   |   -1578.67191557 |  -0.032573   |  0.717409
  28   |   -1578.67182327 |  -0.032481   |  0.717096
  34   |   -1578.66810992 |  -0.028767   |  0.449957
  35   |   -1578.66764499 |  -0.028302   |  0.469805
  37   |   -1578.66540533 |  -0.026063   |  0.556848
  40   |   -1578.66299332 |  -0.023651   |  0.439768
 WARNING!!! Character may have changed, choosing by energy
 ---------------------------------------------------
 Switching to state:          40 from          28
 Trace of A+D (Beta):   1.9186
 Trace of A+D (Alpha):  0.0814
 ---------------------------------------------------
 Stored State:   40 -1578.66299332
 ---------------------------------------------------
 Dynamic thresholds being used
 to change set FOLLOW_ENERGY and FOLLOW_OVERLAP
 ---------------------------------------------------
  SETman timing summary (seconds)
  CPU time            153708.99s
  System time              0.22s
  Wall time             6402.03s

Here it should select:

 21   |   -1578.67768797 |  -0.038345   |  0.825723

with the highest overlap.

The input:

$molecule
   3 2
 Ru                 0.00133100   -0.00306100   -0.00037000
 C                  2.68266900   -1.11322700   -0.56646800
 C                  2.18752300    0.64165700   -2.03124400
 C                  3.40396000    0.49660800   -2.67819600
 C                  4.28145700   -0.48596400   -2.23727800
 C                  3.91936900   -1.29825200   -1.17080400
 C                  2.18777100   -1.91651300    0.55492700
 C                  0.39504400   -2.26304000    2.01607400
 C                  1.06543800   -3.29099800    2.65861700
 C                  2.33723300   -3.63515400    2.21822700
 C                  2.90469500   -2.94323800    1.15626100
 H                  1.46810500    1.39801100   -2.33936100
 H                  3.64967000    1.14912900   -3.51055100
 H                  5.24543700   -0.62307300   -2.72026900
 H                  4.59588700   -2.07029600   -0.81640600
 H                 -0.60257200   -1.95567600    2.32441400
 H                  0.58940600   -3.80633100    3.48729000
 H                  2.88923300   -4.43937500    2.69781500
 H                  3.89862600   -3.20159400    0.80262900
 C                 -2.32028900   -1.74565100   -0.56483800
 C                 -0.55584000   -2.21769600   -2.02451800
 C                 -1.30652000   -3.17637700   -2.68527700
 C                 -2.60648900   -3.41656800   -2.25820900
 C                 -3.11930000   -2.69599200   -1.18741900
 C                 -2.75072300   -0.92238000    0.56877500
 C                 -2.12891100    0.78701500    2.03773500
 C                 -3.34449400    0.71430600    2.69808100
 C                 -4.28575500   -0.21122300    2.26492000
 C                 -3.98797000   -1.03749300    1.18944000
 H                  0.46575600   -1.98998500   -2.32352700
 H                 -0.87057900   -3.71741000   -3.51964300
 H                 -3.22261000   -4.16115900   -2.75551300
 H                 -4.13509000   -2.86986600   -0.84514300
 H                 -1.35974700    1.49458500    2.34178000
 H                 -3.53933300    1.37654500    3.53617800
 H                 -5.24909700   -0.29360500    2.76161900
 H                 -4.71148900   -1.76938600    0.84226400
 C                  0.56917800    2.84253400    0.56577500
 C                  1.09388700    3.97037400    1.18390000
 C                  1.96082300    3.81392300    2.25755200
 C                  2.28992600    2.53585300    2.69161800
 C                  1.74295300    1.44780400    2.03139600
 C                 -0.36500600    2.87745900   -0.56279200
 C                 -0.81175300    4.04281000   -1.17309500
 C                 -1.69462800    3.95260100   -2.24152200
 C                 -2.11591700    2.70247400   -2.67832600
 C                 -1.64116800    1.57559300   -2.02759700
 H                  0.82762500    4.96366300    0.83528500
 H                  2.37502600    4.68894900    2.75172900
 H                  2.96161000    2.37269400    3.52898400
 H                  1.97075500    0.42745200    2.33378600
 H                 -0.47338800    5.01341600   -0.82245300
 H                 -2.04963600    4.85677100   -2.72918200
 H                 -2.80374700    2.59091600   -3.51104400
 H                 -1.94052800    0.57452700   -2.33289500
 N                 -1.84072600   -0.01288300    1.00050500
 N                 -1.05344500   -1.52062800   -0.99315500
 N                  0.94428800   -1.59451000    0.99198200
 N                  1.83762700   -0.14475500   -1.00260700
 N                  0.90379300    1.60033200    0.99681200
 N                 -0.78594400    1.66363000   -0.99864700
$end

$rem
   JOBTYPE             opt
   EXCHANGE            CAM-B3LYP
   CIS_N_ROOTS         40
   CIS_SINGLETS        true
   CIS_TRIPLETS        false
   CIS_STATE_DERIV     28
   STATE_FOLLOW        TRUE
   FOLLOW_ENERGY       1000000
   FOLLOW_OVERLAP      90
   NTO_PAIRS           2
   BASIS               def2-svp
   ECP                 def2-ecp
   XC_GRID             3
   SOLVENT_METHOD      pcm
   THRESH              12
   MEM_TOTAL           50000
   MAX_SCF_CYCLES      1500
   GUI                 2
$end

$pcm
   Theory              CPCM
   Method              SWIG
   Solver              Inversion
   Radii               Bondi
$end

$solvent
   Dielectric          78.39
   OpticalDielectric   1.777849
$end

$van_der_waals
1
   44                   2.02
$end

or here by another set of setting you see that it selects root 38 instead of root 28 with a highest overlap and the same character:

 Step:              1 State:            28
 Previous State: 28  -1578.671124720805
 Estimated State:    -1578.665227032863
 Energy threshold: 0.05898 Overlap threshold:  0.750
 ---------------------------------------------------
 State |    Energy (Eh)   | Delta E (Eh) |  Overlap 
 ---------------------------------------------------
  18   |   -1578.68196179 |  -0.016735   |  0.838127
  20   |   -1578.68112210 |  -0.015895   |  0.806934
  28   |   -1578.67268964 |  -0.007463   |  0.776180
  29   |   -1578.67206946 |  -0.006842   |  0.752917
  30   |   -1578.67175345 |  -0.006526   |  0.748234
  37   |   -1578.66581218 |  -0.000585   |  0.723177
  38   |   -1578.66485578 |   0.000371   |  0.700696
  39   |   -1578.66438689 |   0.000840   |  0.692148
 WARNING!!! Character may have changed, choosing by energy
 ---------------------------------------------------
 Switching to state:          38 from          28
 Trace of A+D (Beta):   1.1040
 Trace of A+D (Alpha):  0.8960
 ---------------------------------------------------
 Stored State:   38 -1578.66485578
 ---------------------------------------------------
 Dynamic thresholds being used
 to change set FOLLOW_ENERGY and FOLLOW_OVERLAP

is there any way to turn off considering Dynamic thresholds?